Human Honeypot
Article URL: https://human2llm.pages.dev/
Comments URL: https://news.ycombinator.com/item?id=49119322
Points: 5
# Comments: 1
Article URL: https://human2llm.pages.dev/
Comments URL: https://news.ycombinator.com/item?id=49119322
Points: 5
# Comments: 1
Hi HN! We’re Akilan and Miguel, the creators of MarbleOS.
The inspiration for Marble comes from the GUI work at Xerox PARC, the 1984 Macintosh, and later NeXTSTEP, which became the foundation for Mac OS X. Before GUIs, interacting with a computer was limited to strange terminal commands like:
C:\> DIR
C:\> COPY FILE.TXT A:
You had to remember the command, syntax, paths, and parameters.
The GUI made those capabilities visible. Instead of remembering commands, you could point at files, drag them, click buttons, and select actions from menus. It did not necessarily make entirely new things possible; it just made existing capabilities much easier to understand and use. We feel like AI is still somewhere around this command-line stage.
Even though the strict syntax has been replaced with natural language, the interaction can still be quite stiff and depend on heavily recall. Tools like Claude Cowork, still look surprisingly terminal-like: /skill-name [param1] [param2].
The parameters are written in natural language, but the user still needs to know which capability exists and how to invoke it. The user has to know what the AI can do, what file to attach, what tool or skill to use, what context it needs, and how to phrase the task.
ChatGPT works very well for asking questions, but we aren't convinced that it's the final interface for delegating work across files, websites, tools, and multiple agents.
Once the AI is actually doing work, a blank text box and a list of chat threads start to feel limiting. Most of the system’s capabilities and most of the task’s structure are invisible. With Marble, we are exploring an interface that treats AI more like a workspace than another chat app. Each delegated task becomes a card. Multiple jobs can sit next to each other and run at once. Files, tools, and finished artifacts are visible at once. Before a task runs, Marble also shows which tools it expects to use. The basic idea is that the user should not have to hold the entire structure of the task in their head. And the result should be something directly usable, like a spreadsheet, PDF, slide deck, or other file, rather than something buried in a transcript.
We’re not claiming to have figured out the final form of GUI for agents. Marble is our attempt to explore whether there is a better interaction model than attaching more tools to a chat box. The product is mainly for people who already use ChatGPT or Claude, but haven’t really adopted agent workflows yet. We’ve found that when the tools and possible actions are made more visible, people start delegating work they would not have thought to do through normal chat.
The site includes a downloadable beta if you want to try it. Does Marble feel like a genuinely novel interface?
Comments URL: https://news.ycombinator.com/item?id=49119274
Points: 4
# Comments: 0
Use a Mattress Firm promo code to save on top mattresses, score a free adjustable base, and unlock up to $300 in instant credits.
Redecorate your home for less with our expert tips on applying SelectBlinds coupons, maximizing holiday sales, and locking in new customer discounts.
Refresh your workout wardrobe. Save on buttery-soft leggings, chic streetwear, and activewear essentials with verified Alo Yoga discount codes, free shipping, and loyalty program perks.
Article URL: https://github.com/openjdk/jdk/pull/31120
Comments URL: https://news.ycombinator.com/item?id=49119063
Points: 15
# Comments: 3
Article URL: https://earendil.com/posts/session-portability/
Comments URL: https://news.ycombinator.com/item?id=49118781
Points: 68
# Comments: 7
Comments URL: https://news.ycombinator.com/item?id=49118696
Points: 108
# Comments: 39
Article URL: https://www.science.org/content/article/california-aquifer-may-have-crossed-point-no-return
Comments URL: https://news.ycombinator.com/item?id=49118663
Points: 121
# Comments: 92
In a review triggered by OpenAI’s Hugging Face incident, Anthropic discovered three of its AI models had breached real organizations during third-party evaluations.